This updates the `Signature` and `CallBinding` machinery to support
multiple overloads for a callable. This is currently only used for
`KnownFunction`s that we special-case in our type inference code. It
does **_not_** yet update the semantic index builder to handle
`@overload` decorators and construct a multi-signature `Overloads`
instance for real Python functions.
While I was here, I updated many of the `try_call` special cases to use
signatures (possibly overloaded ones now) and `bind_call` to check
parameter lists. We still need some of the mutator methods on
`OverloadBinding` for the special cases where we need to update return
types based on some Rust code.
## Summary
One of the motivations in https://github.com/astral-sh/ruff/pull/16428
for panicking when the `test` or `debug_assertions` features are enabled
and a lookup of a `KnownClass` fails is that we've had some latent bugs
in our code where certain variants have been silently falling back to
`Unknown` in every typeshed lookup without us realising. But that in
itself isn't a great motivation for panicking in
`KnownClass::to_instance()`, since we can fairly easily add some tests
that assert that we don't unexpectedly fallback to `Unknown` for any
`KnownClass` variant. This PR adds those tests.
## Test Plan
`cargo test -p red_knot_python_semantic`
## Summary
This mostly fixes#14899
My motivation was similar to the last comment by @sharkdp there. I ran
red_knot on a codebase and the most common error was patterns like this
failing:
```
def foo(x: str): ...
x: Any = ...
if isinstance(x, str):
foo(x) # Object of type `Any & str` cannot be assigned to parameter 1 (`x`) of function `foo`; expected type `str`
```
The desired behavior is pretty much to ignore Any/Unknown when resolving
intersection assignability - `Any & str` should be assignable to `str`,
and `str` should be assignable to `str & Any`
The fix is actually very similar to the existing code in
`is_subtype_of`, we need to correctly handle intersections on either
side, while being careful to handle dynamic types as desired.
This does not fix the second test case from that issue:
```
static_assert(is_assignable_to(Intersection[Unrelated, Any], Not[tuple[Unrelated, Any]]))
```
but that's misleading because the root cause there has nothing to do
with gradual types. I added a simpler test case that also fails:
```
static_assert(is_assignable_to(Unrelated, Not[tuple[Unrelated]]))
```
This is because we don't determine that Unrelated does not subclass from
tuple so we can't rule out this relation. If that logic is improved then
this fix should also handle the case of the intersection
## Test Plan
Added a bunch of is_assignable_to tests, most of which failed before
this fix.
## Summary
Part of https://github.com/astral-sh/ruff/issues/15382
This PR adds support for inferring the `lambda` expression and return
the `CallableType`.
Currently, this is only limited to inferring the parameters and a todo
type for the return type.
For posterity, I tried using the `file_expression_type` to infer the
return type of lambda but it would always lead to cycle. The main reason
is that in `infer_parameter_definition`, the default expression is being
inferred using `file_expression_type`, which is correct, but it then
Take the following source code as an example:
```py
lambda x=1: x
```
Here's how the code will flow:
* `infer_scope_types` for the global scope
* `infer_lambda_expression`
* `infer_expression` for the default value `1`
* `file_expression_type` for the return type using the body expression.
This is because the body creates it's own scope
* `infer_scope_types` (lambda body scope)
* `infer_name_load` for the symbol `x` whose visible binding is the
lambda parameter `x`
* `infer_parameter_definition` for parameter `x`
* `file_expression_type` for the default value `1`
* `infer_scope_types` for the global scope because of the default
expression
This will then reach to `infer_definition` for the parameter `x` again
which then creates the cycle.
## Test Plan
Add tests around `lambda` expression inference.
## Summary
Theoretically this should be slightly more performant, since the
`class.is_known()` calls each do a separate Salsa lookup, which we can
avoid if we do a single `match` on the value of `class.known()`. It also
ends up being two lines less code overall!
## Test Plan
`cargo test -p red_knot_python_semantic`
## Summary
Fixes#16566, fixes#16575
The semantics of `Type::class_member` changed in
https://github.com/astral-sh/ruff/pull/16416, but the property-test
infrastructure was not updated. That means that the property tests were
panicking on the second `expect_type` call here:
0361021863/crates/red_knot_python_semantic/src/types/property_tests.rs (L151-L158)
With the somewhat unhelpful message:
```
Expected a (possibly unbound) type, not an unbound symbol
```
Applying this patch, and then running `QUICKCHECK_TESTS=1000000 cargo
test --release -p red_knot_python_semantic -- --ignored
types::property_tests::stable::equivalent_to_is_reflexive` showed
clearly that it was no longer able to find _any_ methods on _any_
classes due to the change in semantics of `Type::class_member`:
```diff
--- a/crates/red_knot_python_semantic/src/types/property_tests.rs
+++ b/crates/red_knot_python_semantic/src/types/property_tests.rs
@@ -27,7 +27,7 @@
use std::sync::{Arc, Mutex, MutexGuard, OnceLock};
use crate::db::tests::{setup_db, TestDb};
-use crate::symbol::{builtins_symbol, known_module_symbol};
+use crate::symbol::{builtins_symbol, known_module_symbol, Symbol};
use crate::types::{
BoundMethodType, CallableType, IntersectionBuilder, KnownClass, KnownInstanceType,
SubclassOfType, TupleType, Type, UnionType,
@@ -150,10 +150,11 @@ impl Ty {
Ty::BuiltinsFunction(name) => builtins_symbol(db, name).symbol.expect_type(),
Ty::BuiltinsBoundMethod { class, method } => {
let builtins_class = builtins_symbol(db, class).symbol.expect_type();
- let function = builtins_class
- .class_member(db, method.into())
- .symbol
- .expect_type();
+ let Symbol::Type(function, ..) =
+ builtins_class.class_member(db, method.into()).symbol
+ else {
+ panic!("no method `{method}` on class `{class}`");
+ };
create_bound_method(db, function, builtins_class)
}
```
This PR updates the property-test infrastructure to use `Type::member`
rather than `Type::class_member`.
## Test Plan
- Ran `QUICKCHECK_TESTS=1000000 cargo test --release -p
red_knot_python_semantic -- --ignored types::property_tests::stable`
successfully
- Checked that there were no remaining uses of `Type::class_member` in
`property_tests.rs`
## Summary
Fixes a small nit of mine -- we are currently inconsistent in our
spelling between "metaclass" and "meta class", and between "meta type"
and "meta-type". This PR means that we consistently use "metaclass" and
"meta-type".
## Test Plan
`uvx pre-commit run -a`
## Summary
Part of https://github.com/astral-sh/ruff/issues/15382
This PR implements a general callable type that wraps around a
`Signature` and it uses that new type to represent `typing.Callable`.
It also implements `Display` support for `Callable`. The format is as:
```
([<arg name>][: <arg type>][ = <default type>], ...) -> <return type>
```
The `/` and `*` separators are added at the correct boundary for
positional-only and keyword-only parameters. Now, as `typing.Callable`
only has positional-only parameters, the rendered signature would be:
```py
Callable[[int, str], None]
# (int, str, /) -> None
```
The `/` separator represents that all the arguments are positional-only.
The relationship methods that check assignability, subtype relationship,
etc. are not yet implemented and will be done so as a follow-up.
## Test Plan
Add test cases for display support for `Signature` and various mdtest
for `typing.Callable`.
## Summary
Resolves#16365
Add support for unpacking `with` statement targets.
## Test Plan
Added some test cases, alike the ones added by #15058.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
* Attributes/method are now properly looked up on metaclasses, when
called on class objects
* We properly distinguish between data descriptors and non-data
descriptors (but we do not yet support them in store-context, i.e.
`obj.data_descr = …`)
* The descriptor protocol is now implemented in a single unified place
for instances, classes and dunder-calls. Unions and possibly-unbound
symbols are supported in all possible stages of the process by creating
union types as results.
* In general, the handling of "possibly-unbound" symbols has been
improved in a lot of places: meta-class attributes, attributes,
descriptors with possibly-unbound `__get__` methods, instance
attributes, …
* We keep track of type qualifiers in a lot more places. I anticipate
that this will be useful if we import e.g. `Final` symbols from other
modules (see relevant change to typing spec:
https://github.com/python/typing/pull/1937).
* Detection and special-casing of the `typing.Protocol` special form in
order to avoid lots of changes in the test suite due to new `@Todo`
types when looking up attributes on builtin types which have `Protocol`
in their MRO. We previously
looked up attributes in a wrong way, which is why this didn't come up
before.
closes#16367closes#15966
## Context
The way attribute lookup in `Type::member` worked before was simply
wrong (mostly my own fault). The whole instance-attribute lookup should
probably never have been integrated into `Type::member`. And the
`Type::static_member` function that I introduced in my last descriptor
PR was the wrong abstraction. It's kind of fascinating how far this
approach took us, but I am pretty confident that the new approach
proposed here is what we need to model this correctly.
There are three key pieces that are required to implement attribute
lookups:
- **`Type::class_member`**/**`Type::find_in_mro`**: The
`Type::find_in_mro` method that can look up attributes on class bodies
(and corresponding bases). This is a partial function on types, as it
can not be called on instance types like`Type::Instance(…)` or
`Type::IntLiteral(…)`. For this reason, we usually call it through
`Type::class_member`, which is essentially just
`type.to_meta_type().find_in_mro(…)` plus union/intersection handling.
- **`Type::instance_member`**: This new function is basically the
type-level equivalent to `obj.__dict__[name]` when called on
`Type::Instance(…)`. We use this to discover instance attributes such as
those that we see as declarations on class bodies or as (annotated)
assignments to `self.attr` in methods of a class.
- The implementation of the descriptor protocol. It works slightly
different for instances and for class objects, but it can be described
by the general framework:
- Call `type.class_member("attribute")` to look up "attribute" in the
MRO of the meta type of `type`. Call the resulting `Symbol` `meta_attr`
(even if it's unbound).
- Use `meta_attr.class_member("__get__")` to look up `__get__` on the
*meta type* of `meta_attr`. Call it with `__get__(meta_attr, self,
self.to_meta_type())`. If this fails (either the lookup or the call),
just proceed with `meta_attr`. Otherwise, replace `meta_attr` in the
following with the return type of `__get__`. In this step, we also probe
if a `__set__` or `__delete__` method exists and store it in
`meta_attr_kind` (can be either "data descriptor" or "normal attribute
or non-data descriptor").
- Compute a `fallback` type.
- For instances, we use `self.instance_member("attribute")`
- For class objects, we use `class_attr =
self.find_in_mro("attribute")`, and then try to invoke the descriptor
protocol on `class_attr`, i.e. we look up `__get__` on the meta type of
`class_attr` and call it with `__get__(class_attr, None, self)`. This
additional invocation of the descriptor protocol on the fallback type is
one major asymmetry in the otherwise universal descriptor protocol
implementation.
- Finally, we look at `meta_attr`, `meta_attr_kind` and `fallback`, and
handle various cases of (possible) unboundness of these symbols.
- If `meta_attr` is bound and a data descriptor, just return `meta_attr`
- If `meta_attr` is not a data descriptor, and `fallback` is bound, just
return `fallback`
- If `meta_attr` is not a data descriptor, and `fallback` is unbound,
return `meta_attr`
- Return unions of these three possibilities for partially-bound
symbols.
This allows us to handle class objects and instances within the same
framework. There is a minor additional detail where for instances, we do
not allow the fallback type (the instance attribute) to completely
shadow the non-data descriptor. We do this because we (currently) don't
want to pretend that we can statically infer that an instance attribute
is always set.
Dunder method calls can also be embedded into this framework. The only
thing that changes is that *there is no fallback type*. If a dunder
method is called on an instance, we do not fall back to instance
variables. If a dunder method is called on a class object, we only look
it up on the meta class, never on the class itself.
## Test Plan
New Markdown tests.
## Summary
This PR closes#15199.
The change I just made is to set all variables to type `Unknown` if
unpacking fails, but in some cases this may be excessive.
For example:
```py
a, b, c = "ab"
reveal_type(a) # Unknown, but it would be reasonable to think of it as LiteralString
reveal_type(c) # Unknown
```
```py
# Failed to unpack before the starred expression
(a, b, *c, d, e) = (1,)
reveal_type(a) # Unknown
reveal_type(b) # Unknown
...
# Failed to unpack after the starred expression
(a, b, *c, d, e) = (1, 2, 3)
reveal_type(a) # Unknown, but should it be Literal[1]?
reveal_type(b) # Unknown, but should it be Literal[2]?
reveal_type(c) # Todo
reveal_type(d) # Unknown
reveal_type(e) # Unknown
```
I will modify it if you think it would be better to make it a different
type than just `Unknown`.
## Test Plan
I have made appropriate modifications to the test cases affected by this
change, and also added some more test cases.
## Summary
- `Never` is callable
- `Never` is iterable
- Arbitrary attributes can be accessed on `Never`
Split out from #16416 that is going to be required.
## Test Plan
Tests for all properties above.
## Summary
This PR introduces a new mdtest option `system` that can either be
`in-memory` or `os`
where `in-memory` is the default.
The motivation for supporting `os` is so that we can write OS/system
specific tests
with mdtests. Specifically, I want to write mdtests for the module
resolver,
testing that module resolution is case sensitive.
## Test Plan
I tested that the case-sensitive module resolver test start failing when
setting `system = "os"`
## Summary
Python's module resolver is case sensitive.
This PR adds mdtests that assert that our module resolution is case
sensitive.
The tests currently all pass because our in memory file system is case
sensitive.
I'll add support for using the real file system to the mdtest framework
in a separate PR.
This PR also adds support for specifying extra search paths to the
mdtest framework.
## Test Plan
The tests fail when running them using the real file system.
To kick off the work of supporting generics, this adds many new
(currently failing) tests, showing the behavior we plan to support.
This is still missing a lot! Not included:
- typevar tuples
- param specs
- variance
- `Self`
But it's a good start! We can add more failing tests for those once we
tackle these.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
This is split out of https://github.com/astral-sh/ruff/pull/14029, to
reduce the size of that PR, and to validate that this "fallback type"
support in `TypeInference` doesn't come with a performance cost. It also
improves the reliability and debuggability of our current (temporary)
cycle handling.
In order to recover from a cycle, we have to be able to construct a
"default" `TypeInference` where all expressions and definitions have
some "default" type. In our current cycle handling, this "default" type
is just unknown or a todo type. With fixpoint iteration, the "default"
type will be `Type::Never`, which is the "bottom" type that fixpoint
iteration starts from.
Since it would be costly (both in space and time) to actually enumerate
all expressions and definitions in a scope, just to insert the same
default type for all of them, instead we add an optional "missing type"
fallback to `TypeInference`, which (if set) is the fallback type for any
expression or definition which doesn't have an explicit type set.
With this change, cycles can no longer result in the dreaded "Missing
key" errors looking up the type of some expression.
This trait should eventually go away, so we rename it (and supporting
types) to make room for a new concrete `Diagnostic` type.
This commit is just the rename. In the next commit, we'll move it to a
different module.
## Summary
This PR does a small refactor to avoid double
`symbol_table(...).symbol(...)` call to check for `__slots__` and
`TYPE_CHECKING`. It merges them into a single call.
I noticed this while looking at
https://github.com/astral-sh/ruff/pull/16468.
## Summary
This PR adds more features to #16468.
* Adds a new error rule `invalid-type-checking-constant`, which occurs
when we try to assign a value other than `False` to a user-defined
`TYPE_CHECKING` variable (it is possible to assign `...` in a stub
file).
* Allows annotated assignment to `TYPE_CHECKING`. Only types that
`False` can be assigned to are allowed. However, the type of
`TYPE_CHECKING` will be inferred to be `Literal[True]` regardless of
what the type is specified.
## Test plan
I ran the tests with `cargo test -p red_knot_python_semantic` and
confirmed that all tests passed.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
This PR closes#15722.
The change is that if the variable `TYPE_CHECKING` is defined/imported,
the type of the variable is interpreted as `Literal[True]` regardless of
what the value is.
This is compatible with the behavior of other type checkers (e.g. mypy,
pyright).
## Test Plan
I ran the tests with `cargo test -p red_knot_python_semantic` and
confirmed that all tests passed.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
Regardless of whether #16408 and #16311 pan out, this part is worth
pulling out as a separate PR.
Before, you had to define a new `IndexVec` index type for each type of
association list you wanted to create. Now there's a single index type
that's internal to the alist implementation, and you use `List<K, V>` to
store a handle to a particular list.
This also adds some property tests for the alist implementation.
We currently keep two separate pieces of state regarding the current
loop on `SemanticIndexBuilder`. One is an enum simply reflecting whether
we are currently inside a loop, and the other is the saved flow states
for `break` statements found in the current loop.
For adding loopy control flow, I'll need to add some additional loop
state (`continue` states, for example). Prepare for this by
consolidating our existing loop state into a single struct and
simplifying the API for pushing and popping a loop.
This is purely a refactor, so tests are not changed.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Minor follow-up to https://github.com/astral-sh/ruff/pull/16161
This `not_callable` flag wasn't functional, because it could never be
`false`. It was initialized to `true` and then only ever updated with
`|=`, which can never make it `false`.
Add a test that exercises the case where it _should_ be `false` (all of
the union elements are callable) but `bindings` is also empty (all union
elements have binding errors). Before this PR, the added test wrongly
emits a diagnostic that the union `Literal[f1] | Literal[f2]` is not
callable.
And add a test where a union call results in one binding error and one
not-callable error, where we currently give the wrong result (we show
only the binding error), with a TODO.
Also add TODO comments in a couple other tests where ideally we'd report
more than just one error out of a union call.
Also update the flag name to `all_errors_not_callable` to more clearly
indicate the semantics of the flag.
In https://github.com/astral-sh/ruff/pull/16306#discussion_r1966290700,
@carljm pointed out that #16306 introduced a terminology problem, with
too many things called a "constraint". This is a follow-up PR that
renames `Constraint` to `Predicate` to hopefully clear things up a bit.
So now we have that:
- a _predicate_ is a Python expression that might influence type
inference
- a _narrowing constraint_ is a list of predicates that constraint the
type of a binding that is visible at a use
- a _visibility constraint_ is a ternary formula of predicates that
define whether a binding is visible or a statement is reachable
This is a pure renaming, with no behavioral changes.
## Summary
Model dunder-calls correctly (and in one single place), by implementing
this behavior (using `__getitem__` as an example).
```py
def getitem_desugared(obj: object, key: object) -> object:
getitem_callable = find_in_mro(type(obj), "__getitem__")
if hasattr(getitem_callable, "__get__"):
getitem_callable = getitem_callable.__get__(obj, type(obj))
return getitem_callable(key)
```
See the new `calls/dunder.md` test suite for more information. The new
behavior also needs much fewer lines of code (the diff is positive due
to new tests).
## Test Plan
New tests; fix TODOs in existing tests.
This PR adds an implementation of [association
lists](https://en.wikipedia.org/wiki/Association_list), and uses them to
replace the previous `BitSet`/`SmallVec` representation for narrowing
constraints.
An association list is a linked list of key/value pairs. We additionally
guarantee that the elements of an association list are sorted (by their
keys), and that they do not contain any entries with duplicate keys.
Association lists have fallen out of favor in recent decades, since you
often need operations that are inefficient on them. In particular,
looking up a random element by index is O(n), just like a linked list;
and looking up an element by key is also O(n), since you must do a
linear scan of the list to find the matching element. Luckily we don't
need either of those operations for narrowing constraints!
The typical implementation also suffers from poor cache locality and
high memory allocation overhead, since individual list cells are
typically allocated separately from the heap. We solve that last problem
by storing the cells of an association list in an `IndexVec` arena.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
Add a diagnostic if a pure instance variable is accessed on a class object. For example
```py
class C:
instance_only: str
def __init__(self):
self.instance_only = "a"
# error: Attribute `instance_only` can only be accessed on instances, not on the class object `Literal[C]` itself.
C.instance_only
```
---------
Co-authored-by: David Peter <mail@david-peter.de>
## Summary
Add support for `@classmethod`s.
```py
class C:
@classmethod
def f(cls, x: int) -> str:
return "a"
reveal_type(C.f(1)) # revealed: str
```
## Test Plan
New Markdown tests
## Summary
I spotted a minor mistake in my descriptor protocol implementation where
`C.descriptor` would pass the meta type (`type`) of the type of `C`
(`Literal[C]`) as the owner argument to `__get__`, instead of passing
`Literal[C]` directly.
## Test Plan
New test.
Two related changes. For context:
1. We were maintaining two separate arenas of `Constraint`s in each
use-def map. One was used for narrowing constraints, and the other for
visibility constraints. The visibility constraint arena was interned,
ensuring that we always used the same ID for any particular
`Constraint`. The narrowing constraint arena was not interned.
2. The TDD code relies on _all_ TDD nodes being interned and reduced.
This is an important requirement for TDDs to be a canonical form, which
allows us to use a single int comparison to test for "always true/false"
and to compare two TDDs for equivalence. But we also need to support an
individual `Constraint` having multiple values in a TDD evaluation (e.g.
to handle a `while` condition having different values the first time
it's evaluated vs later times). Previously, we handled that by
introducing a "copy" number, which was only there as a disambiguator, to
allow an interned, deduplicated constraint ID to appear in the TDD
formula multiple times.
A better way to handle (2) is to not intern the constraints in the
visibility constraint arena! The caller now gets to decide: if they add
a `Constraint` to the arena more than once, they get distinct
`ScopedConstraintId`s — which the TDD code will treat as distinct
variables, allowing them to take on different values in the ternary
function.
With that in place, we can then consolidate on a single (non-interned)
arena, which is shared for both narrowing and visibility constraints.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
This PR achieves the following:
* Add support for checking method calls, and inferring return types from
method calls. For example:
```py
reveal_type("abcde".find("abc")) # revealed: int
reveal_type("foo".encode(encoding="utf-8")) # revealed: bytes
"abcde".find(123) # error: [invalid-argument-type]
class C:
def f(self) -> int:
pass
reveal_type(C.f) # revealed: <function `f`>
reveal_type(C().f) # revealed: <bound method: `f` of `C`>
C.f() # error: [missing-argument]
reveal_type(C().f()) # revealed: int
```
* Implement the descriptor protocol, i.e. properly call the `__get__`
method when a descriptor object is accessed through a class object or an
instance of a class. For example:
```py
from typing import Literal
class Ten:
def __get__(self, instance: object, owner: type | None = None) ->
Literal[10]:
return 10
class C:
ten: Ten = Ten()
reveal_type(C.ten) # revealed: Literal[10]
reveal_type(C().ten) # revealed: Literal[10]
```
* Add support for member lookup on intersection types.
* Support type inference for `inspect.getattr_static(obj, attr)` calls.
This was mostly used as a debugging tool during development, but seems
more generally useful. It can be used to bypass the descriptor protocol.
For the example above:
```py
from inspect import getattr_static
reveal_type(getattr_static(C, "ten")) # revealed: Ten
```
* Add a new `Type::Callable(…)` variant with the following sub-variants:
* `Type::Callable(CallableType::BoundMethod(…))` — represents bound
method objects, e.g. `C().f` above
* `Type::Callable(CallableType::MethodWrapperDunderGet(…))` — represents
`f.__get__` where `f` is a function
* `Type::Callable(WrapperDescriptorDunderGet)` — represents
`FunctionType.__get__`
* Add new known classes:
* `types.MethodType`
* `types.MethodWrapperType`
* `types.WrapperDescriptorType`
* `builtins.range`
## Performance analysis
On this branch, we do more work. We need to do more call checking, since
we now check all method calls. We also need to do ~twice as many member
lookups, because we need to check if a `__get__` attribute exists on
accessed members.
A brief analysis on `tomllib` shows that we now call `Type::call` 1780
times, compared to 612 calls before.
## Limitations
* Data descriptors are not yet supported, i.e. we do not infer correct
types for descriptor attribute accesses in `Store` context and do not
check writes to descriptor attributes. I felt like this was something
that could be split out as a follow-up without risking a major
architectural change.
* We currently distinguish between `Type::member` (with descriptor
protocol) and `Type::static_member` (without descriptor protocol). The
former corresponds to `obj.attr`, the latter corresponds to
`getattr_static(obj, "attr")`. However, to model some details correctly,
we would also need to distinguish between a static member lookup *with*
and *without* instance variables. The lookup without instance variables
corresponds to `find_name_in_mro`
[here](https://docs.python.org/3/howto/descriptor.html#invocation-from-an-instance).
We currently approximate both using `member_static`, which leads to two
open TODOs. Changing this would be a larger refactoring of
`Type::own_instance_member`, so I chose to leave it out of this PR.
## Test Plan
* New `call/methods.md` test suite for method calls
* New tests in `descriptor_protocol.md`
* New `call/getattr_static.md` test suite for `inspect.getattr_static`
* Various updated tests
## Summary
This avoids looking up `__bool__` on class `bool` for every
`Type::Instance(bool).bool()` call. 1% performance win on cold cache, 4%
win on incremental performance.
This updates the `SymbolBindings` and `SymbolDeclarations` types to use
a single smallvec of live bindings/declarations, instead of splitting
that out into separate containers for each field.
I'm seeing an 11-13% `cargo bench` performance improvement with this
locally (for both cold and incremental). I'm interested to see if
Codspeed agrees!
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
A minor cleanup that breaks up a `HashMap` of an enum into separate
`HashMap`s for each variant. (These separate fields were already how
this cache was being described in the big comment at the top of the
file!)
This is a small tweak to avoid adding the callable `Type` on the error
value itself. Namely, it's always available regardless of the error, and
it's easy to pass it down explicitly to the diagnostic generating code.
It's likely that the other `CallBindingError` variants will also want
the callable `Type` to improve diagnostics too. This way, we don't have
to duplicate the `Type` on each variant. It's just available to all of
them.
Ref https://github.com/astral-sh/ruff/pull/16239#discussion_r1962352646
## Summary
Follow up on the discussion
[here](https://github.com/astral-sh/ruff/pull/16121#discussion_r1962973298).
Replace builtin classes with custom placeholder names, which should
hopefully make the tests a bit easier to understand.
I carefully renamed things one after the other, to make sure that there
is no functional change in the tests.
We now resolve references in "eager" scopes correctly — using the
bindings and declarations that are visible at the point where the eager
scope is created, not the "public" type of the symbol (typically the
bindings visible at the end of the scope).
---------
Co-authored-by: Alex Waygood <alex.waygood@gmail.com>
This uses the refactoring and support for secondary diagnostic messages
to improve the diagnostic for "invalid argument type." The main
improvement here is that we show where the function being called is
defined, and annotate the span corresponding to the invalid parameter.
This is a small little hack to make the `Diagnostic` trait
capable of supporting attaching multiple spans.
This design should be considered transient. This was just the
quickest way that I could see to pass multiple spans through from
the type checker to the diagnostic renderer.
It seems nothing is using it, and I'm not sure if it makes semantic
sense. Particularly if we want to support multiple ranges. One could
make an argument that this ought to correspond to the "primary"
range (which we should have), but I think such a concept is better
expressed as an explicit routine if possible.
## Summary
This PR updates the formatter and linter to use the `PythonVersion`
struct from the `ruff_python_ast` crate internally. While this doesn't
remove the need for the `linter::PythonVersion` enum, it does remove the
`formatter::PythonVersion` enum and limits the use in the linter to
deserializing from CLI arguments and config files and moves most of the
remaining methods to the `ast::PythonVersion` struct.
## Test Plan
Existing tests, with some inputs and outputs updated to reflect the new
(de)serialization format. I think these are test-specific and shouldn't
affect any external (de)serialization.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
This PR does the following:
* Moves the following from `types.rs` in `symbol.rs`:
* `symbol`
* `global_symbol`
* `imported_symbol`
* `symbol_from_bindings`
* `symbol_from_declarations`
* `SymbolAndQualifiers`
* `SymbolFromDeclarationsResult`
* Moves the following from `stdlib.rs` in `symbol.rs` and removes
`stdlib.rs`:
* `known_module_symbol`
* `builtins_symbol`
* `typing_symbol` (only for tests)
* `typing_extensions_symbol`
* `builtins_module_scope`
* `core_module_scope`
* Add `symbol_from_bindings_impl` and `symbol_from_declarations_impl` to
keep `RequiresExplicitReExport` an implementation detail
* Make `declaration_type` a `pub(crate)` as it's required in
`symbol_from_declarations` (`binding_type` is already `pub(crate)`
The main motivation is to keep the implementation details private and
only expose an ergonomic API which uses sane defaults for various
scenario to avoid any mistakes from the caller. Refer to
https://github.com/astral-sh/ruff/pull/16133#discussion_r1955262772,
https://github.com/astral-sh/ruff/pull/16133#issue-2850146612 for
details.
## Summary
This PR makes the following changes:
- It adjusts various callsites to use the new
`ast::StringLiteral::contents_range()` method that was introduced in
https://github.com/astral-sh/ruff/pull/16183. This is less verbose and
more type-safe than using the `ast::str::raw_contents()` helper
function.
- It adds a new `ast::ExprStringLiteral::as_unconcatenated_literal()`
helper method, and adjusts various callsites to use it. This addresses
@MichaReiser's review comment at
https://github.com/astral-sh/ruff/pull/16183#discussion_r1957334365.
There is no functional change here, but it helps readability to make it
clearer that we're differentiating between implicitly concatenated
strings and unconcatenated strings at various points.
- It renames the `StringLiteralValue::flags()` method to
`StringLiteralFlags::first_literal_flags()`. If you're dealing with an
implicitly concatenated string `string_node`,
`string_node.value.flags().closer_len()` could give an incorrect result;
this renaming makes it clearer that the `StringLiteralFlags` instance
returned by the method is only guaranteed to give accurate information
for the first `StringLiteral` contained in the `ExprStringLiteral` node.
- It deletes the unused `BytesLiteralValue::flags()` method. This seems
prone to misuse in the same way as `StringLiteralValue::flags()`: if
it's an implicitly concatenated bytestring, the `BytesLiteralFlags`
instance returned by the method would only give accurate information for
the first `BytesLiteral` in the bytestring.
## Test Plan
`cargo test`
## Summary
Running `cargo test -p red_knot_python_semantic` failed because of a
missing serde feature. This PR enables the `ruff_python_ast`'`s `serde`
if the crate's `serde` feature is enabled
## Test Plan
`cargo test -p red_knot_python_semantic` compiles again
When adjusting the existing tests, I aimed to avoid dealing with the
special case in other tests if it's not necessary to do so (that is,
avoid using `float` and `complex` as examples where we just need "some
type"), and keep the tests for the special case mostly collected in the
mdtest dedicated to that purpose.
Fixes https://github.com/astral-sh/ruff/issues/14932
## Summary
This PR moves the `PythonVersion` struct from the
`red_knot_python_semantic` crate to the `ruff_python_ast` crate so that
it can be used more easily in the syntax error detection work. Compared
to that [prototype](https://github.com/astral-sh/ruff/pull/16090/) these
changes reduce us from 2 `PythonVersion` structs to 1.
This does not unify any of the `PythonVersion` *enums*, but I hope to
make some progress on that in a follow-up.
## Test Plan
Existing tests, this should not change any external behavior.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
This PR refactors the symbol lookup APIs to better facilitate the
re-export implementation. Specifically,
* Add `module_type_symbol` which returns the `Symbol` that's a member of
`types.ModuleType`
* Rename `symbol` -> `symbol_impl`; add `symbol` which delegates to
`symbol_impl` with `RequireExplicitReExport::No`
* Update `global_symbol` to do `symbol_impl` -> fall back to
`module_type_symbol` and default to `RequireExplicitReExport::No`
* Add `imported_symbol` to do `symbol_impl` with
`RequireExplicitReExport` as `Yes` if the module is in a stub file else
`No`
* Update `known_module_symbol` to use `imported_symbol` with a fallback
to `module_type_symbol`
* Update `ModuleLiteralType::member` to use `imported_symbol` with a
custom fallback
We could potentially also update `symbol_from_declarations` and
`symbol_from_bindings` to avoid passing in the `RequireExplicitReExport`
as it would be always `No` if called directly. We could add
`symbol_from_declarations_impl` and `symbol_from_bindings_impl`.
Looking at the `_impl` functions, I think we should move all of these
symbol related logic into `symbol.rs` where `Symbol` is defined and the
`_impl` could be private while we expose the public APIs at the crate
level. This would also make the `RequireExplicitReExport` an
implementation detail and the caller doesn't need to worry about it.
This is an alternative implementation to #15848.
## Summary
This PR adds support for re-export conventions for imports for stub
files.
**How does this work?**
* Add a new flag on the `Import` and `ImportFrom` definitions to
indicate whether they're being exported or not
* Add a new enum to indicate whether the symbol lookup is happening
within the same file or is being queried from another file (e.g., an
import statement)
* When a `Symbol` is being queried, we'll skip the definitions that are
(a) coming from a stub file (b) external lookup and (c) check the
re-export flag on the definition
This implementation does not yet support `__all__` and `*` imports as
both are features that needs to be implemented independently.
closes: #14099closes: #15476
## Test Plan
Add test cases, update existing ones if required.
## Summary
After I was asked twice within the same day, I thought it would be a
good idea to write some *user facing* documentation that explains our
reasoning behind inferring `Unknown | T_inferred` for public uses of
undeclared symbols. This is a major deviation from the behavior of other
type checkers and it seems like a good practice to defend our choice
like this.
This essentially makes it impossible to construct a `Diagnostic`
that has a `TextRange` but no `File`.
This is meant to be a precursor to multi-span support.
(Note that I consider this more of a prototyping-change and not
necessarily what this is going to look like longer term.)
Reviewers can probably review this PR as one big diff instead of
commit-by-commit.
## Summary
This is a follow up to
https://github.com/astral-sh/ruff/pull/15763#discussion_r1949681336
It reverts the change to using ptr equality for `AstNodeRef`s, which in
turn removes the `Eq`, `PartialEq`, and `Hash` implementations for
`AstNodeRef`s parametrized with AST nodes.
Cheap comparisons shouldn't be needed because the node field is
generally marked as `[#tracked]` and `#[no_eq]` and removing the
implementations even enforces that those
attributes are set on all `AstNodeRef` fields (which is good).
The only downside this has is that we technically wouldn't have to mark
the `Unpack::target` as `#[tracked]` because
the `target` field is accessed in every query accepting `Unpack` as an
argument.
Overall, enforcing the use of `#[tracked]` seems like a good trade off,
espacially considering that it's very likely that
we'd probably forget to mark the `Unpack::target` field as tracked if we
add a new `Unpack` query that doesn't access the target.
## Test Plan
`cargo test`
## Summary
Transition to using coarse-grained tracked structs (depends on
https://github.com/salsa-rs/salsa/pull/657). For now, this PR doesn't
add any `#[tracked]` fields, meaning that any changes cause the entire
struct to be invalidated. It also changes `AstNodeRef` to be
compared/hashed by pointer address, instead of performing a deep AST
comparison.
## Test Plan
This yields a 10-15% improvement on my machine (though weirdly some runs
were 5-10% without being flagged as inconsistent by criterion, is there
some non-determinism involved?). It's possible that some of this is
unrelated, I'll try applying the patch to the current salsa version to
make sure.
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
- Simplify unions with `object` to `object`.
- Add a new `Type::object(db)` constructor to abbreviate
`KnownClass::Object.to_instance(db)` in some places.
- Add a `Type::is_object` and `Class::is_object` function to make some
tests for a bit easier to read.
closes#16084
## Test Plan
New Markdown tests.
## Summary
This PR adds support for user-level configurations
(`~/.config/knot/knot.toml`) to Red Knot.
Red Knot will watch the user-level configuration file for changes but
only if it exists
when the process start. It doesn't watch for new configurations,
mainly to simplify things for now (it would require watching the entire
`.config` directory because the `knot` subfolder might not exist
either).
The new `ConfigurationFile` struct seems a bit overkill for now but I
plan to use it for
hierarchical configurations as well.
Red Knot uses the same strategy as uv and Ruff by using the etcetera
crate.
## Test Plan
Added CLI and file watching test
## Summary
This PR adds a new `user_configuration_directory` method to `System`. We
need it to resolve where to lookup a user-level `knot.toml`
configuration file.
The method belongs to `System` because not all platforms have a
convention of where to store such configuration files (e.g. wasm).
I refactored `TestSystem` to be a simple wrapper around an `Arc<dyn
System...>` and use the `System.as_any` method instead to cast it down
to an `InMemory` system. I also removed some `System` specific methods
from `InMemoryFileSystem`, they don't belong there.
This PR removes the `os` feature as a default feature from `ruff_db`.
Most crates depending on `ruff_db` don't need it because they only
depend on `System` or only depend on `os` for testing. This was
necessary to fix a compile error with `red_knot_wasm`
## Test Plan
I'll make use of the method in my next PR. So I guess we won't know if
it works before then but I copied the code from Ruff/uv, so I have high
confidence that it is correct.
`cargo test`
## Summary
This PR generalize the idea that we may want to emit diagnostics for
invalid or incompatible configuration values similar to how we already
do it for `rules`.
This PR introduces a new `Settings` struct that is similar to `Options`
but, unlike
`Options`, are fields have their default values filled in and they use a
representation optimized for reads.
The diagnostics created during loading the `Settings` are stored on the
`Project` so that we can emit them when calling `check`.
The motivation for this work is that it simplifies adding new settings.
That's also why I went ahead and added the `terminal.error-on-warning`
setting to demonstrate how new settings are added.
## Test Plan
Existing tests, new CLI test.
## Summary
No functional change here; this is another simplification split out from
my outcome-refactor branch to reduce the diff there. This merges
`TypeInferenceBuilder::infer_name_load` and
`TypeInferenceBuilder::lookup_name`. This removes the need to have
extensive doc-comments about the purpose of
`TypeInferenceBuilder::lookup_name`, since the method only makes sense
when called from the specific context of
`TypeInferenceBuilder::infer_name_load`.
## Test Plan
`cargo test -p red_knot_python_semantic`
## Summary
- Do not return `Option<Type<…>>` from `Unpacker::get`, but just `Type`.
Panic otherwise.
- Rename `Unpacker::get` to `Unpacker::expression_type`
## Summary
* Support assignments to attributes in more cases:
- assignments in `for` loops
- in unpacking assignments
* Add test for multi-target assignments
* Add tests for all other possible assignments to attributes that could
possibly occur (in decreasing order of likeliness):
- augmented attribute assignments
- attribute assignments in `with` statements
- attribute assignments in comprehensions
- Note: assignments to attributes in named expressions are not
syntactically allowed
closes#15962
## Test Plan
New Markdown tests
## Summary
This PR reverts the behavior changes from
https://github.com/astral-sh/ruff/pull/15990
But it isn't just a revert, it also:
* Adds a test covering this specific behavior
* Preserves the improvement to use `saturating_sub` in the package case
to avoid overflows in the case of invalid syntax
* Use `ancestors` instead of a `for` loop
## Test Plan
Added test
## Summary
Adds a JSON schema generation step for Red Knot. This PR doesn't yet add
a publishing step because it's still a bit early for that
## Test plan
I tested the schema in Zed, VS Code and PyCharm:
* PyCharm: You have to manually add a schema mapping (settings JSON
Schema Mappings)
* Zed and VS code support the inline schema specification
```toml
#:schema /Users/micha/astral/ruff/knot.schema.json
[environment]
extra-paths = []
[rules]
call-possibly-unbound-method = "error"
unknown-rule = "error"
# duplicate-base = "error"
```
```json
{
"$schema": "file:///Users/micha/astral/ruff/knot.schema.json",
"environment": {
"python-version": "3.13",
"python-platform": "linux2"
},
"rules": {
"unknown-rule": "error"
}
}
```
https://github.com/user-attachments/assets/a18fcd96-7cbe-4110-985b-9f1935584411
The Schema overall works but all editors have their own quirks:
* PyCharm: Hovering a name always shows the section description instead
of the description of the specific setting. But it's the same for other
settings in `pyproject.toml` files 🤷
* VS Code (JSON): Using the generated schema in a JSON file gives
exactly the experience I want
* VS Code (TOML):
* Properties with multiple possible values are repeated during
auto-completion without giving any hint how they're different. 
* The property description mushes together the description of the
property and the value, which looks sort of ridiculous. 
* Autocompletion and documentation hovering works (except the
limitations mentioned above)
* Zed:
* Very similar to VS Code with the exception that it uses the
description attribute to distinguish settings with multiple possible
values 
I don't think there's much we can do here other than hope (or help)
editors improve their auto completion. The same short comings also apply
to ruff, so this isn't something new. For now, I think this is good
enough
## Summary
I noticed that the diagnostic range in specific unpacking assignments is
wrong. For this example
```py
a, b = 1
```
we previously got (see first commit):
```
error: lint:not-iterable
--> /src/mdtest_snippet.py:1:1
|
1 | a, b = 1
| ^^^^ Object of type `Literal[1]` is not iterable
|
```
and with this change, we get:
```
error: lint:not-iterable
--> /src/mdtest_snippet.py:1:8
|
1 | a, b = 1
| ^ Object of type `Literal[1]` is not iterable
|
```
## Test Plan
New snapshot tests.
## Summary
Fixes https://github.com/astral-sh/ruff/issues/15989
Red Knot failed to resolve relative imports if the importing module is
located at a search path root.
The issue was that the module resolver returned an `Err(TooManyDots)` as
soon as the parent of the current module is `None` (which is the case
for a module at the search path root).
However, this is incorrect if a `tail` (a module name) exists.
## Summary
- Minor wording update
- Code improvement (thanks Alex)
- Removed all unnecessary filenames throughout our Markdown tests (two
new ones were added in the meantime)
- Minor rewording of the statically-known-branches introduction
This example from @sharkdp shows how terminal statements can appear in
statically known branches:
https://github.com/astral-sh/ruff/pull/15676#issuecomment-2618809716
```py
def _(cond: bool):
x = "a"
if cond:
x = "b"
if True:
return
reveal_type(x) # revealed: "a", "b"; should be "a"
```
We now use visibility constraints to track reachability, which allows us
to model this correctly. There are two related changes as a result:
- New bindings are not assumed to be visible; they inherit the current
"scope start" visibility, which effectively means that new bindings are
visible if/when the current flow is reachable
- When simplifying visibility constraints after branching control flow,
we only simplify if none of the intervening branches included a terminal
statement. That is, earlier unaffected bindings are only _actually_
unaffected if all branches make it to the merge point.
## Summary
Allow for literate style in Markdown tests and merge multiple (unnamed)
code blocks into a single embedded file.
closes#15941
## Test Plan
- Interactively made sure that error-lines were reported correctly in
multi-snippet sections.
This causes the diagnostic to highlight the actual unresovable import
instead of the entire `from ... import ...` statement.
While we're here, we expand the test coverage to cover all of the
possible ways that an `import` or a `from ... import` can fail.
Some considerations:
* The first commit in this PR adds a regression test for the current
behavior.
* This creates a new `mdtest/diagnostics` directory. Are folks cool
with this? I guess the idea is to put tests more devoted to diagnostics
than semantics in this directory. (Although I'm guessing there will
be some overlap.)
Fixes#15866
## Summary
This is a first step towards creating a test suite for
[descriptors](https://docs.python.org/3/howto/descriptor.html). It does
not (yet) aim to be exhaustive.
relevant ticket: #15966
## Test Plan
Compared desired behavior with the runtime behavior and the behavior of
existing type checkers.
---------
Co-authored-by: Mike Perlov <mishamsk@gmail.com>
This ties together everything from the previous commits.
Some interesting bits here are how the snapshot is generated
(where we include relevant info to make it easier to review
the snapshots) and also a tweak to how inline assertions are
processed.
This commit also includes some example snapshots just to get
a sense of what they look like. Follow-up work should add
more of these I think.
This makes it possible for callers to set where snapshots
should be stored. In general, I think we expect this to
always be set, since otherwise snapshots will end up in
`red_knot_test`, which is where the tests are actually run.
But that's overall counter-intuitive. This permits us to
store snapshots from mdtests alongside the mdtests themselves.
## Summary
This PR adds `Type::call_bound` method for calls that should follow
descriptor protocol calling convention. The PR is intentionally shallow
in scope and only fixes#15672
Couple of obvious things that weren't done:
* Switch to `call_bound` everywhere it should be used
* Address the fact, that red_knot resolves `__bool__ = bool` as a Union,
which includes `Type::Dynamic` and hence fails to infer that the
truthiness is always false for such a class (I've added a todo comment
in mdtests)
* Doesn't try to invent a new type for descriptors, although I have a
gut feeling it may be more convenient in the end, instead of doing
method lookup each time like I did in `call_bound`
## Test Plan
* extended mdtests with 2 examples from the issue
* cargo neatest run
We now use ternary decision diagrams (TDDs) to represent visibility
constraints. A TDD is just like a BDD ([_binary_ decision
diagram](https://en.wikipedia.org/wiki/Binary_decision_diagram)), but
with "ambiguous" as an additional allowed value. Unlike the previous
representation, TDDs are strongly normalizing, so equivalent ternary
formulas are represented by exactly the same graph node, and can be
compared for equality in constant time.
We currently have a slight 1-3% performance regression with this in
place, according to local testing. However, we also have a _5× increase_
in performance for pathological cases, since we can now remove the
recursion limit when we evaluate visibility constraints.
As follow-on work, we are now closer to being able to remove the
`simplify_visibility_constraint` calls in the semantic index builder. In
the vast majority of cases, we now see (for instance) that the
visibility constraint after an `if` statement, for bindings of symbols
that weren't rebound in any branch, simplifies back to `true`. But there
are still some cases we generate constraints that are cyclic. With
fixed-point cycle support in salsa, or with some careful analysis of the
still-failing cases, we might be able to remove those.
## Summary
I experimented with [not trimming trailing newlines in code
snippets](https://github.com/astral-sh/ruff/pull/15926#discussion_r1940992090),
but since came to the conclusion that the current behavior is better
because otherwise, there is no way to write snippets without a trailing
newline at all. And when you copy the code from a Markdown snippet in
GitHub, you also don't get a trailing newline.
I was surprised to see some test failures when I played with this
though, and decided to make this test independent from this
implementation detail.
## Summary
This is a follow-up to #15726, #15778, and #15794 to preserve the triple
quote and prefix flags in plain strings, bytestrings, and f-strings.
I also added a `StringLiteralFlags::without_triple_quotes` method to
avoid passing along triple quotes in rules like SIM905 where it might
not make sense, as discussed
[here](https://github.com/astral-sh/ruff/pull/15726#discussion_r1930532426).
## Test Plan
Existing tests, plus many new cases in the `generator::tests::quote`
test that should cover all combinations of quotes and prefixes, at least
for simple string bodies.
Closes#7799 when combined with #15694, #15726, #15778, and #15794.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
Resolves#15695, rework of #15704.
This change modifies the Mdtests framework so that:
* Paths must now be specified in a separate preceding line:
`````markdown
`a.py`:
```py
x = 1
```
`````
If the path of a file conflicts with its `lang`, an error will be
thrown.
* Configs are no longer accepted. The pattern still take them into
account, however, to avoid "Unterminated code block" errors.
* Unnamed files are now assigned unique, `lang`-respecting paths
automatically.
Additionally, all legacy usages have been updated.
## Test Plan
Unit tests and Markdown tests.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
This extracts some pure refactoring noise from
https://github.com/astral-sh/ruff/pull/15861. This changes the API for
creating and evaluating visibility constraints, but does not change how
they are respresented internally. There should be no behavioral or
performance changes in this PR.
Changes:
- Hide the internal representation isn't changed, so that we can make
changes to it in #15861.
- Add a separate builder type for visibility constraints. (With TDDs, we
will have some additional builder state that we can throw away once
we're done constructing.)
- Remove a layer of helper methods from `UseDefMapBuilder`, making
`SemanticIndexBuilder` responsible for constructing whatever visibility
constraints it needs.
## Summary
Add support for implicitly-defined instance attributes, i.e. support
type inference for cases like this:
```py
class C:
def __init__(self) -> None:
self.x: int = 1
self.y = None
reveal_type(C().x) # int
reveal_type(C().y) # Unknown | None
```
## Benchmarks
Codspeed reports no change in a cold-cache benchmark, and a -1%
regression in the incremental benchmark. On `black`'s `src` folder, I
don't see a statistically significant difference between the branches:
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:---|---:|---:|---:|---:|
| `./red_knot_main check --project /home/shark/black/src` | 133.7 ± 9.5 | 126.7 | 164.7 | 1.01 ± 0.08 |
| `./red_knot_feature check --project /home/shark/black/src` | 132.2 ± 5.1 | 118.1 | 140.9 | 1.00 |
## Test Plan
Updated and new Markdown tests
## Summary
Related to #15848, this PR adds the imports explicitly as we'll now flag
these symbols as undefined.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
This mimics a simplification we have on the OR side, where we simplify
`A ∨ !A` to true. This requires changes to how we add `while` statements
to the semantic index, since we now need distinct
`VisibilityConstraint`s if we need to model evaluating a `Constraint`
multiple times at different points in the execution of the program.
Something Alex and I threw together during our 1:1 this morning. Allows
us to collect statistics on the prevalence of various types in a file,
most usefully TODO types or other dynamic types.
`FlowSnapshot` now tracks a `reachable` bool, which indicates whether we
have encountered a terminal statement on that control flow path. When
merging flow states together, we skip any that have been marked
unreachable. This ensures that bindings that can only be reached through
unreachable paths are not considered visible.
## Test Plan
The new mdtests failed (with incorrect `reveal_type` results, and
spurious `possibly-unresolved-reference` errors) before adding the new
visibility constraints.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
When we discussed the plan on how to proceed with instance attributes,
we said that we should first extend our research into the behavior of
existing type checkers. The result of this research is summarized in the
newly added / modified tests in this PR. The TODO comments align with
existing behavior of other type checkers. If we deviate from the
behavior, it is described in a comment.
## Summary
Adds a slightly more comprehensive documentation of our behavior
regarding type inference for public uses of symbols. In particular:
- What public type do we infer for `x: int = any()`?
- What public type do we infer for `x: Unknown = 1`?
## Summary
On `main`, red-knot:
- Considers `P | Q` equivalent to `Q | P`
- Considered `tuple[P | Q]` equivalent to `tuple[Q | P]`
- Considers `tuple[P | tuple[P | Q]]` equivalent to `tuple[tuple[Q | P]
| P]`
- ‼️ Does _not_ consider `tuple[tuple[P | Q]]` equivalent to
`tuple[tuple[Q | P]]`
The key difference for the last one of these is that the union appears
inside a tuple that is directly nested inside another tuple.
This PR fixes this so that differently ordered unions are considered
equivalent even when they appear inside arbitrarily nested tuple types.
## Test Plan
- Added mdtests that fails on `main`
- Checked that all property tests continue to pass with this PR
This is a follow-up to #15702 that hopefully claws back the 1%
performance regression. Assuming it works, the trick is to iterate over
the constraints vectors via mut reference (aka a single pointer), so
that we're not copying `BitSet`s into and out of the zip tuples as we
iterate. We use `std::mem::take` as a poor-man's move constructor only
at the very end, when we're ready to emplace it into the result. (C++
idioms intended! 😄)
With local testing via hyperfine, I'm seeing this be 1-3% faster than
`main` most of the time — though a small number of runs (1 in 10,
maybe?) are a wash or have `main` faster. Codspeed reports a 2%
gain.
## Summary
Use `Unknown | T_inferred` as the type for *undeclared* public symbols.
## Test Plan
- Updated existing tests
- New test for external `__slots__` modifications.
- New tests for external modifications of public symbols.
## Summary
Another small PR to focus #15674 solely on the relevant changes. This
makes our Markdown tests less dependent on precise types of public
symbols, without actually changing anything semantically in these tests.
Best reviewed using ignore-whitespace-mode.
## Test Plan
Tested these changes on `main` and on the branch from #15674.
## Summary
Make the remaining `infer.rs` unit tests independent from public symbol
type inference decisions (see upcoming change in #15674).
## Test Plan
- Made sure that the unit tests actually fail if one of the
`assert_type` assertions is changed.
## Summary
Port comprehension tests from Rust to Markdown
I don' think the remaining tests in `infer.rs` should be ported to
Markdown, maybe except for the incremental-checking tests when (if ever)
we have support for that in the MD tests.
closes#13696
## Summary
- Port "deferred annotations" unit tests to Markdown
- Port `implicit_global_in_function` unit test to Markdown
- Removed `resolve_method` and `local_inference` unit tests. These seem
like relics from a time where type inference was in it's early stages.
There is no way that these tests would fail today without lots of other
things going wrong as well.
part of #13696
based on #15683
## Test Plan
New MD tests for existing Rust unit tests.
## Summary
- Add feature to specify a custom typeshed from within Markdown-based
tests
- Port "builtins" unit tests from `infer.rs` to Markdown tests, part of
#13696
## Test Plan
- Tests for the custom typeshed feature
- New Markdown tests for deleted Rust unit tests
## Summary
Somehow, I managed to crash the `mdtest` runner today. I struggled to
reproduce this again to see if it's actually fixed (even with an
artificial `sleep` between the two `cargo test` invocations), but the
original backtrace clearly showed that this is where the problem
originated from. And it seems like a clear TOCTOU problem.
## Summary
Raise "invalid-assignment" diagnostics for incorrect assignments to
attributes, for example:
```py
class C:
var: str = "a"
C.var = 1 # error: "Object of type `Literal[1]` is not assignable to `str`"
```
closes#15456
## Test Plan
- Updated test assertions
- New test for assignments to module-attributes
## Summary
This PR generalizes some of the logic we have in `Type::is_subtype_of`
and `Type::is_disjoint_from` so that we fallback to the instance type of
the metaclass more often in `Type::ClassLiteral` and `Type::SubclassOf`
branches. This simplifies the code (we end up with one less branch in
`is_subtype_of`, and we can remove a helper method that's no longer
used), makes the code more robust (any fixes made to subtyping or
disjointness of instance types will automatically improve our
understanding of subtyping/disjointness for class-literal types and
`type[]` types) and more elegantly expresses the type-system invariants
encoded in these branches.
## Test Plan
No new tests added (it's a pure refactor, adding no new functionality).
All existing tests pass, however, including the property tests.
The AST generator creates a reference enum for each syntax group — an
enum where each variant contains a reference to the relevant syntax
node. Previously you could customize the name of the reference enum for
a group — primarily because there was an existing `ExpressionRef` type
that wouldn't have lined up with the auto-derived name `ExprRef`. This
follow-up PR is a simple search/replace to switch over to the
auto-derived name, so that we can remove this customization point.
## Summary
Test executables usually write failure messages (including panics) to
stdout, but I just managed to make a mdtest crash with
```
thread 'mdtest__unary_not' has overflowed its stack
fatal runtime error: stack overflow
```
which is printed to stderr. This test simply appends stderr to stdout
(`stderr=subprocess.STDOUT` can not be used with `capture_output`)
## Test Plan
Make sure that the error message is now visible in the output of `uv -q
run crates/red_knot_python_semantic/mdtest.py`
## Summary
The `Options` struct is intended to capture the user's configuration
options but
`EnvironmentOptions::venv_path` supports both a `SitePackages::Known`
and `SitePackages::Derived`.
Users should only be able to provide `SitePackages::Derived`—they
specify a path to a venv, and Red Knot derives the path to the
site-packages directory. We'll only use the `Known` variant once we
automatically discover the Python installation.
That's why this PR changes `EnvironmentOptions::venv_path` from
`Option<SitePackages>` to `Option<SystemPathBuf>`.
This requires making some changes to the file watcher test, and I
decided to use `extra_paths` over venv path
because our venv validation is annoyingly correct -- making mocking a
venv rather involved.
## Test Plan
`cargo test`
## Summary
As more and more tests move to Markdown, running the mdtest suite
becomes one of the most common tasks for developers working on Red Knot.
There are a few pain points when doing so, however:
- The `build.rs` script enforces recompilation (~five seconds) whenever
something changes in the `resource/mdtest` folder. This is strictly
necessary, because whenever files are added or removed, the test harness
needs to be updated. But this is very rarely the case! The most common
scenario is that a Markdown file has *changed*, and in this case, no
recompilation is necessary. It is currently not possible to distinguish
these two cases using `cargo::rerun-if-changed`. One can work around
this by running the test executable manually, but it requires finding
the path to the correct `mdtest-<random-hash>` executable.
- All Markdown tests are run by default. This is needed whenever Rust
code changes, but while working on the tests themselves, it is often
much more convenient to only run the tests for a single file. This can
be done by using a `mdtest__path_to_file` filter, but this needs to be
manually spelled out or copied from the test output.
- `cargo`s test output for a failing Markdown test is often
unnecessarily verbose. Unless there is an *actual* panic somewhere in
the code, mdtests usually fail with the explicit *"Some tests failed"*
panic in the mdtest suite. But in those cases, we are not interested in
the pointer to the source of this panic, but only in the mdtest suite
output.
This PR adds a Markdown test runner tool that attempts to make the
developer experience better.
Once it is started using
```bash
uv run -q crates/red_knot_python_semantic/mdtest.py
```
it will first recompile the tests once (if cargo requires it), find the
path to the `mdtest` executable, and then enter into a mode where it
watches for changes in the `red_knot_python_semantic` crate. Whenever …
* … a Markdown file changes, it will rerun the mdtest for this specific
file automatically (no recompilation!).
* … a Markdown file is added, it will recompile the tests and then run
the mdtest for the new file
* … Rust code is changed, it will recompile the tests and run all of
them
The tool also trims down `cargo test` output and only shows the actual
mdtest errors.
The tool will certainly require a few more iterations before it becomes
mature, but I'm curious to hear if there is any interest for something
like this.
## Test Plan
- Tested the new runner under various scenarios.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
Rename two functions with outdated names (they used to return `Type`s):
* `bindings_ty` => `symbol_from_bindings` (returns `Symbol`)
* `declarations_ty` => `symbol_from_declarations` (returns a
`SymbolAndQualifiers` result)
I chose `symbol_from_*` instead of `*_symbol` as I found the previous
name quite confusing. Especially since `binding_ty` and `declaration_ty`
also exist (singular).
## Test Plan
—
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
Add support for `typing.ClassVar`, i.e. emit a diagnostic in this
scenario:
```py
from typing import ClassVar
class C:
x: ClassVar[int] = 1
c = C()
c.x = 3 # error: "Cannot assign to pure class variable `x` from an instance of type `C`"
```
## Test Plan
- New tests for the `typing.ClassVar` qualifier
- Fixed one TODO in `attributes.md`
## Summary
This is a small, tentative step towards the bigger goal of understanding
instance attributes.
- Adds partial support for pure instance variables declared in the class
body, i.e. this case:
```py
class C:
variable1: str = "a"
variable2 = "b"
reveal_type(C().variable1) # str
reveal_type(C().variable2) # Unknown | Literal["b"]
```
- Adds `property` as a known class to query for `@property` decorators
- Splits up various `@Todo(instance attributes)` cases into
sub-categories.
## Test Plan
Modified existing MD tests.
## Summary
This PR adds support for configuring Red Knot in the `tool.knot` section
of the project's
`pyproject.toml` section. Options specified on the CLI precede the
options in the configuration file.
This PR only supports the `environment` and the `src.root` options for
now.
Other options will be added as separate PRs.
There are also a few concerns that I intentionally ignored as part of
this PR:
* Handling of relative paths: We need to anchor paths relative to the
current working directory (CLI), or the project (`pyproject.toml` or
`knot.toml`)
* Tracking the source of a value. Diagnostics would benefit from knowing
from which configuration a value comes so that we can point the user to
the right configuration file (or CLI) if the configuration is invalid.
* Schema generation and there's a lot more; see
https://github.com/astral-sh/ruff/issues/15491
This PR changes the default for first party codes: Our existing default
was to only add the project root. Now, Red Knot adds the project root
and `src` (if such a directory exists).
Theoretically, we'd have to add a file watcher event that changes the
first-party search paths if a user later creates a `src` directory. I
think this is pretty uncommon, which is why I ignored the complexity for
now but I can be persuaded to handle it if it's considered important.
Part of https://github.com/astral-sh/ruff/issues/15491
## Test Plan
Existing tests, new file watching test demonstrating that changing the
python version and platform is correctly reflected.
## Summary
Closes https://github.com/astral-sh/ruff/issues/15508
For any two instance types `T` and `S`, we know they are disjoint if
either `T` is final and `T` is not a subclass of `S` or `S` is final and
`S` is not a subclass of `T`.
Correspondingly, for any two types `type[T]` and `S` where `S` is an
instance type, `type[T]` can be said to be disjoint from `S` if `S` is
disjoint from `U`, where `U` is the type that represents all instances
of `T`'s metaclass.
And a heterogeneous tuple type can be said to be disjoint from an
instance type if the instance type is disjoint from `tuple` (a type
representing all instances of the `tuple` class at runtime).
## Test Plan
- A new mdtest added. Most of our `is_disjoint_from()` tests are not
written as mdtests just yet, but it's pretty hard to test some of these
edge cases from a Rust unit test!
- Ran `QUICKCHECK_TESTS=1000000 cargo test --release -p
red_knot_python_semantic -- --ignored types::property_tests::stable`
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
Ref: https://github.com/astral-sh/ruff/pull/15387#discussion_r1917796907
This PR updates `F722` to show syntax error message instead of the
string content.
I think it's more useful to show the syntax error message than the
string content. In the future, when the diagnostics renderer is more
capable, we could even highlight the exact location of the syntax error
along with the annotation string.
This is also in line with how we show the diagnostic in red knot.
## Test Plan
Update existing test snapshots.
## Summary
Resolves#9467
Parse quoted annotations as if the string content is inside parenthesis.
With this logic `x` and `y` in this example are equal:
```python
y: """
int |
str
"""
z: """(
int |
str
)
"""
```
Also this rule only applies to triple
quotes([link](https://github.com/python/typing-council/issues/9#issuecomment-1890808610)).
This PR is based on the
[comments](https://github.com/astral-sh/ruff/issues/9467#issuecomment-2579180991)
on the issue.
I did one extra change, since we don't want any indentation tokens I am
setting the `State::Other` as the initial state of the Lexer.
Remaining work:
- [x] Add a test case for red-knot.
- [x] Add more tests.
## Test Plan
Added a test which previously failed because quoted annotation contained
indentation.
Added an mdtest for red-knot.
Updated previous test.
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
If `S <: T`, then `~T <: ~S`. This test currently fails with example
like:
```
S = tuple[()]
T = ~Literal[True] & ~Literal[False]
```
`T` is equivalent to `~(Literal[True] | Literal[False])` and therefore
equivalent to `~bool`, but the minimal example for a failure is what is
stated above. We correctly recognize that `S <: T`, but fail to see that
`~T <: ~S`, i.e. `bool <: ~tuple[()]`.
This is why the tests goes into the "flaky" section as well.
## Test Plan
```
export QUICKCHECK_TESTS=100000
while cargo test --release -p red_knot_python_semantic -- --ignored types::property_tests::flaky::negation_reverses_subtype_order; do :; done
```
## Summary
Adds some initial tests for class and instance attributes, mostly to
document (and discuss) what we want to support eventually. These
tests are not exhaustive yet. The idea is to specify the coarse-grained
behavior first.
Things that we'll eventually want to test:
- Interplay with inheritance
- Support `Final` in addition to `ClassVar`
- Specific tests for `ClassVar`, like making sure that we support things
like `x: Annotated[ClassVar[int], "metadata"]`
- … or making sure that we raise an error here:
```py
class Foo:
def __init__(self):
self.x: ClassVar[str] = "x"
```
- Add tests for `__new__` in addition to the tests for `__init__`
- Add tests that show that we use the union of types if multiple methods
define the symbol with different types
- Make sure that diagnostics are raised if, e.g., the inferred type of
an assignment within a method does not match the declared type in the
class body.
- https://github.com/astral-sh/ruff/pull/15474#discussion_r1916556284
- Method calls are completely left out for now.
- Same for `@property`
- … and the descriptor protocol
## Test Plan
New Markdown tests
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
The next sync of typeshed would have failed without manual changes
anyway, so I'm doing one manual sync + the required changes in our
`sys.platform` tests (which are necessary because of my tiny typeshed PR
here: https://github.com/python/typeshed/pull/13378).
closes#15485 (the next run of the pipeline in two weeks should be fine
as the bug has been fixed upstream)
## Summary
Adds two additional tests for `is_equivalent_to` so that we cover all
properties of an [equivalence relation].
## Test Plan
```
while cargo test --release -p red_knot_python_semantic -- --ignored types::property_tests::stable; do :; done
```
[equivalence relation]:
https://en.wikipedia.org/wiki/Equivalence_relation
## Summary
This changeset adds new tests for public uses of symbols,
considering all possible declaredness and boundness states.
Note that this is a mere documentation of the current behavior. There is
still an [open ticket] questioning some of these choices (or unintential
behaviors).
## Test plan
Made sure that the respective test fails if I add the questionable case
again in `symbol_by_id`:
```rs
Symbol::Type(inferred_ty, Boundness::Bound) => {
Symbol::Type(inferred_ty, Boundness::Bound)
}
```
[open ticket]: https://github.com/astral-sh/ruff/issues/14297
## Summary
Simplification follow-up to #15413.
There's no need to have a dedicated `CallOutcome` variant for every
known function, it's only necessary if the special-cased behavior of the
known function includes emitting extra diagnostics. For `typing.cast`,
there's no such need; we can use the regular `Callable` outcome variant,
and update the return type according to the cast. (This is the same way
we already handle `len`.)
One reason to avoid proliferating unnecessary `CallOutcome` variants is
that currently we have to explicitly add emitting call-binding
diagnostics, for each outcome variant. So we were previously wrongly
silencing any binding diagnostics on calls to `typing.cast`. Fixing this
revealed a separate bug, that we were emitting a bogus error anytime
more than one keyword argument mapped to a `**kwargs` parameter. So this
PR also adds test and fix for that bug.
## Test Plan
Existing `cast` tests pass unchanged, added new test for `**kwargs` bug.
## Summary
In `SymbolState` merging, use `BitSet::union` instead of inserting
declarations one by one. This used to be the case but was changed in
https://github.com/astral-sh/ruff/pull/15019 because we had to iterate
over declarations anyway.
This is an alternative to https://github.com/astral-sh/ruff/pull/15419
by @MichaReiser. It's similar in performance, but a bit more
declarative and less imperative.
## Summary
Follow-up PR from https://github.com/astral-sh/ruff/pull/15415🥲
The exact same property test already exists:
`intersection_assignable_to_both` and
`all_type_pairs_can_be_assigned_from_their_intersection`
## Test Plan
`cargo test -p red_knot_python_semantic -- --ignored
types::property_tests::flaky`
A small PR to reduce some of the code duplication between the various
branches, make it a little more readable and move the API closer to what
we already have for `KnownClass`
## Summary
[**Rendered version of the new test
suite**](https://github.com/astral-sh/ruff/blob/david/intersection-type-tests/crates/red_knot_python_semantic/resources/mdtest/intersection_types.md)
Moves most of our existing intersection-types tests to a dedicated
Markdown test suite, extends the test coverage, unifies the notation for
these tests, groups tests into a proper structure, and adds some
explanations for various simplification strategies.
This changeset also:
- Adds a new simplification where `~Never` is removed from
intersections.
- Adds a new simplification where adding `~object` simplifies the whole
intersection to `Never`
- Avoids unnecessary assignment-checks between inferred and declared
type. This was added to this changeset to avoid many false positive
errors in this test suite.
Resolves the task described in this old comment
[here](e01da82a5a..e7e432bca2 (r1819924085)).
## Test Plan
Running the new Markdown tests
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Prompted by
> One nit: I think we need to consider `Any` and `Unknown` and `Todo` as
all (gradually) equivalent to each other, and thus `type & Any` and
`type & Unknown` and `type & Todo` as also equivalent. The distinction
between `Any` vs `Unknown` vs `Todo` is entirely about
provenance/debugging, there is no type level distinction. (And I've been
wondering if the `Any` vs `Unknown` distinction is really worth it.)
The thought here is that _most_ places want to treat `Any`, `Unknown`,
and `Todo` identically. So this PR simplifies things by having a single
`Type::Any` variant, and moves the provenance part into a new `AnyType`
type. If you need to treat e.g. `Todo` differently, you still can by
pattern-matching into the `AnyType`. But if you don't, you can just use
`Type::Any(_)`.
(This would also allow us to (more easily) distinguish "unknown via an
unannotated value" from "unknown because of a typing error" should we
want to do that in the future)
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Co-authored-by: Carl Meyer <carl@astral.sh>
## Summary
This moves almost all of our existing `UnionBuilder` tests to a
Markdown-based test suite.
I see how this could be a more controversial change, since these tests
where written specifically for `UnionBuilder`, and by creating the union
types using Python type expressions, we add an additional layer on top
(parsing and inference of these expressions) that moves these tests away
from clean unit tests more in the direction of integration tests. Also,
there are probably a few implementation details of `UnionBuilder` hidden
in the test assertions (e.g. order of union elements after
simplifications).
That said, I think we would like to see all those properties that are
being tested here from *any* implementation of union types. And the
Markdown tests come with the usual advantages:
- More consice
- Better readability
- No re-compiliation when working on tests
- Easier to add additional explanations and structure to the test suite
This changeset adds a few additional tests, but keeps the logic of the
existing tests except for a few minor modifications for consistency.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Co-authored-by: T-256 <132141463+T-256@users.noreply.github.com>
## Summary
- Add a workflow to run property tests on a daily basis (based on
`daily_fuzz.yaml`)
- Mark `assignable_to_is_reflexive` as flaky (related to #14899)
- Add new (failing) `intersection_assignable_to_both` test (also related
to #14899)
## Test Plan
Ran:
```bash
export QUICKCHECK_TESTS=100000
while cargo test --release -p red_knot_python_semantic -- \
--ignored types::property_tests::stable; do :; done
```
Observed successful property_tests CI run
## Summary
This changeset migrates all existing `is_assignable_to` tests to a
Markdown-based test. It also increases our test coverage in a hopefully
meaningful way (not claiming to be complete in any sense). But at least
I found and fixed one bug while doing so.
## Test Plan
Ran property tests to make sure the new test succeeds after fixing it.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
Adds a type-check-time Python API that allows us to create and
manipulate types and to test various of their properties. For example,
this can be used to write a Markdown test to make sure that `A & B` is a
subtype of `A` and `B`, but not of an unrelated class `C` (something
that requires quite a bit more code to do in Rust):
```py
from knot_extensions import Intersection, is_subtype_of, static_assert
class A: ...
class B: ...
type AB = Intersection[A, B]
static_assert(is_subtype_of(AB, A))
static_assert(is_subtype_of(AB, B))
class C: ...
static_assert(not is_subtype_of(AB, C))
```
I think this functionality is also helpful for interactive debugging
sessions, in order to query various properties of Red Knot's type
system. Which is something that otherwise requires a custom Rust unit
test, some boilerplate code and constant re-compilation.
## Test Plan
- New Markdown tests
- Tested the modified typeshed_sync workflow locally
## Summary
`Type[Any]` should be assignable to `object`. All types should be
assignable to `object`.
We specifically didn't understand the former; this PR adds a test for
it, and a case to ensure that `Type[Any]` is assignable to anything that
`type` is assignable to (which includes `object`).
This PR also adds a property test that all types are assignable to
object. In order to make it pass, I added a special case to check early
if we are assigning to `object` and just return `true`. In principle,
once we get all the more general cases correct, this special case might
be removable. But having the special case for now allows the property
test to pass.
And we add a property test that all types are subtypes of object. This
failed for the case of an intersection with no positive elements (that
is, a negation type). This really does need to be a special case for
`object`, because there is no other type we can know that a negation
type is a subtype of.
## Test Plan
Added unit test and property test.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
We now support class patterns in a match statement, adding a narrowing
constraint that within the body of that match arm, we can assume that
the subject is an instance of that class.
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
This implements checking of calls.
I ended up following Micha's original suggestion from back when the
signature representation was first introduced, and flattening it to a
single array of parameters. This turned out to be easier to manage,
because we can represent parameters using indices into that array, and
represent the bound argument types as an array of the same length.
Starred and double-starred arguments are still TODO; these won't be very
useful until we have generics.
The handling of diagnostics is just hacked into `return_ty_result`,
which was already inconsistent about whether it emitted diagnostics or
not; now it's even more inconsistent. This needs to be addressed, but
could be a follow-up.
The new benchmark errors here surface the need for intersection support
in `is_assignable_to`.
Fixes#14161.
## Test Plan
Added mdtests.
## Summary
When debugging, I frequently want to know which symbols are being looked
up. `symbol_by_id` adds tracing information, but it only shows the
`ScopedSymbolId`. Since `symbol_by_id` is only called from `symbol`, it
seems reasonable to move the tracing call one level up from
`symbol_by_id` to `symbol`, where we can also show the name of the
symbol.
**Before**:
```
6 └─┐red_knot_python_semantic::types::infer::infer_expression_types{expression=Id(60de), file=/home/shark/tomllib_modified/_parser.py}
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(33)}
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(123)}
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(54)}
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(122)}
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(165)}
6 ┌─┘
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(32)}
6 ┌─┘
6 └─┐red_knot_python_semantic::types::symbol_by_id{symbol=ScopedSymbolId(232)}
6 ┌─┘
6 ┌─┘
6 ┌─┘
6┌─┘
```
**After**:
```
5 └─┐red_knot_python_semantic::types::infer::infer_expression_types{expression=Id(60de), file=/home/shark/tomllib_modified/_parser.py}
5 └─┐red_knot_python_semantic::types::symbol{name="dict"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="dict"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="list"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="list"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="isinstance"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="isinstance"}
5 ┌─┘
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="ValueError"}
5 ┌─┘
5 └─┐red_knot_python_semantic::types::symbol{name="ValueError"}
5 ┌─┘
5 ┌─┘
5 ┌─┘
5┌─┘
```
## Test Plan
```
cargo run --bin red_knot -- --current-directory path/to/tomllib -vvv
```
## Summary
While looking at #14899, I looked at seeing if I could get shrinking on
the examples. It turned out to be straightforward, with a couple of
caveats.
I'm calling `clone` a lot during shrinking. Since by the shrink step
we're already looking at a test failure this feels fine? Unless I
misunderstood `quickcheck`'s core loop
When shrinking `Intersection`s, in order to just rely on `quickcheck`'s
`Vec` shrinking without thinking about it too much, the shrinking
strategy is:
- try to shrink the negative side (keeping the positive side the same)
- try to shrink the positive side (keeping the negative side the same)
This means that you can't shrink from `(A & B & ~C & ~D)` directly to
`(A & ~C)`! You would first need an intermediate failure at `(A & B &
~C)` or `(A & ~C & ~D)`. This feels good enough. Shrinking the negative
side first also has the benefit of trying to strip down negative
elements in these intersections.
## Test Plan
`cargo test -p red_knot_python_semantic -- --ignored
types::property_tests::stable` still fails as it current does on `main`,
but now the errors seem more minimal.
Just like in #15045 for unary expressions: In binary expressions, we
were only looking for dunder expressions for `Type::Instance` types. We
had some special cases for coercing the various `Literal` types into
their corresponding `Instance` types before doing the lookup. But we can
side-step all of that by using the existing `Type::to_meta_type` and
`Type::to_instance` methods.
Resolves#14840
## Summary
Usage of ellipsis literal as default argument is allowed in stub files.
## Test Plan
Added mdtest for both python files and stub files.
---------
Co-authored-by: Carl Meyer <carl@oddbird.net>
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
The test expression in an `elif` clause is evaluated whether or not we
take the branch. Our control flow model for if/elif chains failed to
reflect this, causing wrong inference in cases where an assignment
expression occurs inside an `elif` test expression. Our "no branch taken
yet" snapshot (which is the starting state for every new elif branch)
can't simply be the pre-if state, it must be updated after visiting each
test expression.
Once we do this, it also means we no longer need to track a vector of
narrowing constraints to reapply for each new branch, since our "branch
not taken" state (which is the initial state for each branch) is
continuously updated to include the negative narrowing constraints of
all previous branches.
Fixes#15033.
## Test Plan
Added mdtests.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
We understand `sys.version_info` branches now! As such, I _believe_ this
branch is no longer required; all tests pass without it. I also ran
`QUICKCHECK_TESTS=100000 cargo test -p red_knot_python_semantic --
--ignored types::property_tests::stable`, and no tests failed except for
the known issue with `Type::is_assignable_to()`
(https://github.com/astral-sh/ruff/issues/14899)
## Test Plan
See above
## Summary
Remove `Type::tuple` in favor of `TupleType::from_elements`, avoid a few
intermediate `Vec`tors. Resolves an old [review
comment](https://github.com/astral-sh/ruff/pull/14744#discussion_r1867493706).
## Test Plan
New regression test for something I ran into while implementing this.
## Summary
Part of #13773
This PR adds diagnostics when there is a length mismatch during
unpacking between the number of target expressions and the number of
types for the unpack value expression.
There are 3 cases of diagnostics here where the first two occurs when
there isn't a starred expression and the last one occurs when there's a
starred expression:
1. Number of target expressions is **less** than the number of types
that needs to be unpacked
2. Number of target expressions is **greater** then the number of types
that needs to be unpacked
3. When there's a starred expression as one of the target expression and
the number of target expressions is greater than the number of types
Examples for all each of the above cases:
```py
# red-knot: Too many values to unpack (expected 2, got 3) [lint:invalid-assignment]
a, b = (1, 2, 3)
# red-knot: Not enough values to unpack (expected 2, got 1) [lint:invalid-assignment]
a, b = (1,)
# red-knot: Not enough values to unpack (expected 3 or more, got 2) [lint:invalid-assignment]
a, *b, c, d = (1, 2)
```
The (3) case is a bit special because it uses a distinct wording
"expected n or more" instead of "expected n" because of the starred
expression.
### Location
The diagnostic location is the target expression that's being unpacked.
For nested targets, the location will be the nested expression. For
example:
```py
(a, (b, c), d) = (1, (2, 3, 4), 5)
# ^^^^^^
# red-knot: Too many values to unpack (expected 2, got 3) [lint:invalid-assignment]
```
For future improvements, it would be useful to show the context for why
this unpacking failed. For example, for why the expected number of
targets is `n`, we can highlight the relevant elements for the value
expression.
In the **ecosystem**, **Pyright** uses the target expressions for
location while **mypy** uses the value expression for the location. For
example:
```py
if 1:
# mypy: Too many values to unpack (2 expected, 3 provided) [misc]
# vvvvvvvvv
a, b = (1, 2, 3)
# ^^^^
# Pyright: Expression with type "tuple[Literal[1], Literal[2], Literal[3]]" cannot be assigned to target tuple
# Type "tuple[Literal[1], Literal[2], Literal[3]]" is incompatible with target tuple
# Tuple size mismatch; expected 2 but received 3 [reportAssignmentType]
# red-knot: Too many values to unpack (expected 2, got 3) [lint:invalid-assignment]
```
## Test Plan
Update existing test cases TODO with the error directives.
## Summary
Ref:
3533d7f5b4 (r150651102)
This PR removes the `Ranged` implementation on `DefinitionKind` and
instead uses a method called `target_range` to avoid any confusion about
what range this is for i.e., it's not the range of the node that
represents the definition.
## Summary
Related to #13773
This PR adds support for unpacking `for` statement targets.
This involves updating the `value` field in the `Unpack` target to use
an enum which specifies the "where did the value expression came from?".
This is because for an iterable expression, we need to unpack the
iterator type while for assignment statement we need to unpack the value
type itself. And, this needs to be done in the unpack query.
### Question
One of the ways unpacking works in `for` statement is by looking at the
union of the types because if the iterable expression is a tuple then
the iterator type will be union of all the types in the tuple. This
means that the test cases that will test the unpacking in `for`
statement will also implicitly test the unpacking union logic. I was
wondering if it makes sense to merge these cases and only add the ones
that are specific to the union unpacking or for statement unpacking
logic.
## Test Plan
Add test cases involving iterating over a tuple type. I've intentionally
left out certain cases for now and I'm curious to know any thoughts on
the above query.
## Summary
This changeset adds support for precise type-inference and
boundness-handling of definitions inside control-flow branches with
statically-known conditions, i.e. test-expressions whose truthiness we
can unambiguously infer as *always false* or *always true*.
This branch also includes:
- `sys.platform` support
- statically-known branches handling for Boolean expressions and while
loops
- new `target-version` requirements in some Markdown tests which were
now required due to the understanding of `sys.version_info` branches.
closes#12700closes#15034
## Performance
### `tomllib`, -7%, needs to resolve one additional module (sys)
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:---|---:|---:|---:|---:|
| `./red_knot_main --project /home/shark/tomllib` | 22.2 ± 1.3 | 19.1 |
25.6 | 1.00 |
| `./red_knot_feature --project /home/shark/tomllib` | 23.8 ± 1.6 | 20.8
| 28.6 | 1.07 ± 0.09 |
### `black`, -6%
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:---|---:|---:|---:|---:|
| `./red_knot_main --project /home/shark/black` | 129.3 ± 5.1 | 119.0 |
137.8 | 1.00 |
| `./red_knot_feature --project /home/shark/black` | 136.5 ± 6.8 | 123.8
| 147.5 | 1.06 ± 0.07 |
## Test Plan
- New Markdown tests for the main feature in
`statically-known-branches.md`
- New Markdown tests for `sys.platform`
- Adapted tests for `EllipsisType`, `Never`, etc
## Summary
Refer:
https://github.com/astral-sh/ruff/issues/13773#issuecomment-2548020368
This PR adds support for unpacking union types.
Unpacking a union type requires us to first distribute the types for all
the targets that are involved in an unpacking. For example, if there are
two targets and a union type that needs to be unpacked, each target will
get a type from each element in the union type.
For example, if the type is `tuple[int, int] | tuple[int, str]` and the
target has two elements `(a, b)`, then
* The type of `a` will be a union of `int` and `int` which are at index
0 in the first and second tuple respectively which resolves to an `int`.
* Similarly, the type of `b` will be a union of `int` and `str` which
are at index 1 in the first and second tuple respectively which will be
`int | str`.
### Refactors
There are couple of refactors that are added in this PR:
* Add a `debug_assertion` to validate that the unpack target is a list
or a tuple
* Add a separate method to handle starred expression
## Test Plan
Update `unpacking.md` with additional test cases that uses union types.
This is done using parameter type hints style.
## Summary
This PR adds initial support for `type: ignore`. It doesn't do anything
fancy yet like:
* Detecting invalid type ignore comments
* Detecting type ignore comments that are part of another suppression
comment: `# fmt: skip # type: ignore`
* Suppressing specific lints `type: ignore [code]`
* Detecting unsused type ignore comments
* ...
The goal is to add this functionality in separate PRs.
## Test Plan
---------
Co-authored-by: Carl Meyer <carl@astral.sh>
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
We have a handy `to_meta_type` that does the right thing for class
instances, and also works for all of the other types that are “instances
of” something. Unless I'm missing something, this should let us get rid
of the catch-all clause in one fell swoop.
cf #14548
## Summary
I'm currently on the fence about landing the #14760 PR because it's
unclear how we'd support tracking used and unused suppression comments
in a performant way:
* Salsa adds an "untracked" dependency to every query reading
accumulated values. This has the effect that the query re-runs on every
revision. For example, a possible future query
`unused_suppression_comments(db, file)` would re-run on every
incremental change and for every file. I don't expect the operation
itself to be expensive, but it all adds up in a project with 100k+ files
* Salsa collects the accumulated values by traversing the entire query
dependency graph. It can skip over sub-graphs if it is known that they
contain no accumulated values. This makes accumulators a great tool for
when they are rare; diagnostics are a good example. Unfortunately,
suppressions are more common, and they often appear in many different
files, making the "skip over subgraphs" optimization less effective.
Because of that, I want to wait to adopt salsa accumulators for type
check diagnostics (we could start using them for other diagnostics)
until we have very specific reasons that justify regressing incremental
check performance.
This PR does a "small" refactor that brings us closer to what I have in
#14760 but without using accumulators. To emit a diagnostic, a method
needs:
* Access to the db
* Access to the currently checked file
This PR introduces a new `InferContext` that holds on to the db, the
current file, and the reported diagnostics. It replaces the
`TypeCheckDiagnosticsBuilder`. We pass the `InferContext` instead of the
`db` to methods that *might* emit diagnostics. This simplifies some of
the `Outcome` methods, which can now be called with a context instead of
a `db` and the diagnostics builder. Having the `db` and the file on a
single type like this would also be useful when using accumulators.
This PR doesn't solve the issue that the `Outcome` types feel somewhat
complicated nor that it can be annoying when you need to report a
`Diagnostic,` but you don't have access to an `InferContext` (or the
file). However, I also believe that accumulators won't solve these
problems because:
* Even with accumulators, it's necessary to have a reference to the file
that's being checked. The struggle would be to get a reference to that
file rather than getting a reference to `InferContext`.
* Users of the `HasTy` trait (e.g., a linter) don't want to bother
getting the `File` when calling `Type::return_ty` because they aren't
interested in the created diagnostics. They just want to know what
calling the current expression would return (and if it even is a
callable). This is what the different methods of `Outcome` enable today.
I can ask for the return type without needing extra data that's only
relevant for emitting a diagnostic.
A shortcoming of this approach is that it is now a bit confusing when to
pass `db` and when an `InferContext`. An option is that we'd make the
`file` on `InferContext` optional (it won't collect any diagnostics if
`None`) and change all methods on `Type` to take `InferContext` as the
first argument instead of a `db`. I'm interested in your opinion on
this.
Accumulators are definitely harder to use incorrectly because they
remove the need to merge the diagnostics explicitly and there's no risk
that we accidentally merge the diagnostics twice, resulting in
duplicated diagnostics. I still value performance more over making our
life slightly easier.
Douglas Creager
Alex Waygood
Joey Bar
Dhruv Manilawala
David Peter
Eric Mark Martin
Shunsuke Shibayama
Micha Reiser
Carl Meyer
Andrew Gallant
Mike Perlov
Brent Westbrook
Ibraheem Ahmed
InSync
wooly18
Shaygan Hooshyari
cake-monotone
Raphael Gaschignard
Shantanu